A blueprint for banks going AI-first

Insights

• Banks must go AI-first to create new revenue streams and keep up with digital challengers in the market.
• Agentic AI is playing a key role here; to get ahead, leaders must ensure their workforce are guardians of the AI systems they oversee, ensuring AI’s reliability, quality, security, and ethical decision-making.
• But going AI-first is not that simple. Core processes need to be changed, and AI used as a key part of strategic decision making.
• Change starts with small-scale AI projects, a cross-functional ethics committee, and continuous learning and feedback loops built into product delivery: a digital factory can help.
• This then sets the stage for a more inclusive, adaptive, and forward-looking banking future — a live AI-first enterprise that continuously evolves and learns as new knowledge comes to light.

Over the past two decades, banks have been investing in digital and cloud transformation and are already modernizing their IT landscapes. This has led them to become data-rich and digitally automated. Now, with all the advancements in artificial intelligence (AI) and generative AI, banks can build on this digital, data, and cloud foundation to weave in AI capabilities across the value chain and go AI-first.

By employing AI in everything from knowledge management and communication, information retrieval, fraud detection, reporting, customer sentiment analytics, and credit review processes, AI-first banks can achieve higher margins, create new revenue streams and business models, design better products, accrue higher valuations, and become more productive.

Financial services ahead of the curve

Gartner forecasts that by 2026, businesses will create 60% of new website and mobile app code using generative AI. It will enable the automation of up to 70% of business activities across most occupations by 2030, while also giving employees the chance to take on higher-value work.

Since the launch of ChatGPT by OpenAI in November 2022, significant developments have happened. We now have many large language models, both closed access and open access. Over the last year, we have seen generative AI applied to several business domains — sales and marketing, risk and compliance, customer service, contact centers, business and IT operations, software engineering, employee experience, and learning.

$114 billion was spent on AI by fintech companies in 2022-2023 alone, according to the State of AI annual report. This makes the industry second only to healthcare in the number of AI deals made. According to the Infosys Bank Index Volume 4 report, generative AI and machine learning (ML) spending is expected to increase by 7.7% this year, and AI and cybersecurity now account for 54% of quarterly bank tech budgets.

Wall Street giants, including J.P. Morgan, Bank of America, and others, globally spent over $20 billion on AI in 2023 alone, a number likely to significantly grow in the coming years.

The big questions for financial services executives and the wider organization are: How can we leverage the power of AI to leapfrog in the AI era? What factors will lead to success, and what will diminish its uptake in the organization? And, what will a bank look like that uses AI and data throughout all operations and customer experience?

Continuous learning

However, going AI-first at enterprise scale is not that simple.

There are six key components that will be affected by AI-first banking, starting with strategic decision-making and operations and right down to culture and human resources. The conceptual blueprint in Figure 1 involves more than just using AI tools, however; it signifies a fundamental shift in how a bank operates, thinks, and plans for the future.

Figure 1. Six key components in AI-first banking

Source: Infosys

a. Strategic decision-making and business operations

AI is set to revolutionize strategic decision-making and business operations across three key areas:

1. Core decision-making: Predictive analytics will forecast market trends, customer behavior, and potential risks, enabling proactive and informed strategic decisions. AI's capability in scenario modeling allows for the simulation of various business scenarios, providing leaders with predictive and data-driven insights to guide their planning and decision-making.
2. Data-driven insights: Through customer data analysis, AI helps understand customer needs and preferences, allowing banks to shape business strategies that align with customer expectations.
3. Continuous innovation and adaptation: AI helps design financial products and, by analyzing customer feedback and market trends, drives continuous improvement of services.

b. Customer experience and service enhancement

From personalized investment advice to predictive account management, AI offers banks a more nuanced and responsive customer experience, fostering deeper relationships.

AI-based banking products can also use synthetic customers. These are human avatars with personality and knowledge who can interact with prospects. Compared to an FAQ or chatbot, synthetic customers are based on design personas; they have a story, a goal, and a unique relationship with the bank. These avatars leverage both factual knowledge and personality traits of the customer base and can be used to create proposals, demonstrate product and service value, and assist employees.

c. Operational efficiency and automation

The latest advances in AI will enable agentic systems, one of our top-10 AI imperatives for 2025. Use cases range from complex task execution to project management execution (Figure 2).

Figure 2. Agentic AI

Source: Infosys

At the core of this transformation is process automation: Agentic systems are playing a pivotal role in automating transaction processing, reducing processing times and error rates in routine banking transactions. This not only leads to heightened operational efficiency but also boosts customer satisfaction.

AI algorithms can also analyze vast amounts of operational data, pinpointing areas for improvement, optimizing resource allocation, and streamlining processes. This fosters a data-centric decision-making culture within the bank. Furthermore, AI is being used for predictive maintenance, anticipating equipment and system failures, and enabling preemptive maintenance. This reduces downtime and ensures uninterrupted banking services.

d. Risk management and compliance

Advanced risk assessment ensures stability and integrity of financial institutions. In the AI-first bank, credit risk analysis scrutinizes customer data to assess creditworthiness, minimizing default risk. Similarly, market risk management leverages AI to predict market trends, aiding in managing market-related risks more effectively.

In fraud prevention, real-time fraud monitoring utilizes AI to continuously scan transactions and identify patterns that traditional methods might miss. Anomaly detection strengthens this by pinpointing unusual activities for investigation. Compliance monitoring and reporting has also evolved, with AI systems keeping pace with regulatory changes to ensure compliance and reduce legal risks. Automated processes in efficient reporting reduce human error, streamlining compliance. AI's role in anti-money laundering is crucial, analyzing transactions for potential money laundering indicators and enhancing customer due diligence for regulatory adherence.

e. Security and data privacy

Protecting customer data and ensuring privacy are not just regulatory requirements but central to maintaining trust and integrity. Product innovation must be balanced with a strong security and privacy-by-design culture.

Threat detection and response takes a front seat with AI systems monitoring network traffic and user behavior. Complementing this, automated vulnerability assessments ensure the bank's digital infrastructure remains robust against attacks, with AI constantly scanning for and addressing vulnerabilities.

AI shores up the protection of data via AI-driven access control systems that analyze behavioral patterns, while privacy-preserving AI techniques such as federated learning (i.e., a well-rounded education for computer algorithms without the need for a centralized classroom) become instrumental, enabling analytics while keeping customer data encrypted and private.

f. Organizational culture and human resources

Becoming an AI-first entity marks a shift in an organization’s ethos, placing innovation, adaptability, and continuous learning at the heart of its culture.

The AI-first bank should train and adapt its workforce to work effectively with AI systems. People need different skills in the AI era. The idea of teaming has changed; people will now not only work with each other but also with AI systems. There will be a stronger focus on conflict resolution, building trust, and recognition of value created. Leadership also plays a key role. Leaders are not mere participants in the AI-first transformation, but the driving force behind it.

“Good leaders leverage AI for organizational growth, informed decision-making, and setting visionary, data-driven directions,” says Rafee Tarafdar, CTO at Infosys. “In the AI-powered operating model, leaders are innovators, designers, and integrators. They are also guardians, ensuring the reliability, quality, security, and ethical decision-making of the AI systems they oversee.”

The challenges of change

According to a survey Infosys conducted at an annual Infosys AI event in Boston in September 2024, change management is a significant bottleneck to scaling AI. In fact, it is even more of a concern to executives than budget and data readiness (Figure 3).

Figure 3. Change management is a key challenge

Source: Infosys Topaz, September 2024

The best way to move forward in building the AI-first bank is to start small using a micro-is-the-new-mega approach to change management. This bite-sized approach prioritizes high-impact internal areas, which can then be scaled externally only once significant ROI is achieved. AI products are piloted on a small scale, and lessons from that pilot are used to refine and then scale the rollout, whether for a single point solution (a Q&A chatbot, say), an internal development platform, or a full-scale customer experience AI platform.

Six guiding principles are important:

1. Start with small-scale AI projects: Starting small within an AI foundry — in essence, a digital factory for AI — enables continuous innovation and the industrialization of AI processes across the organization.
2. Build a cross-functional AI ethics committee: Establish a committee responsible for overseeing the ethical aspects of AI deployment.
3. Engage stakeholders early and often: Regularly communicate with stakeholders, including employees, customers, and regulators, about the bank's AI initiatives. Transparency in AI development and deployment is crucial for building trust and ensuring that stakeholders’ concerns are addressed.
4. Incorporate continuous learning and feedback loops: As AI systems are deployed, continuously collect feedback from users and stakeholders. This feedback should be used to make iterative improvements to AI systems, ensuring they remain aligned with ethical standards and organizational goals.
5. Focus on incremental improvements: This approach allows for more manageable and responsible AI integration, reducing risks and allowing for fine-tuning based on feedback and performance.
6. Align AI initiatives with organizational values: Ensure that AI initiatives are in line with the bank’s core values and mission. This alignment helps maintain a consistent approach to ethics and governance across the organization.

The emphasis on micro changes allows for consideration of the ethical implications of AI, ensuring that the bank's transition to AI-driven operations is both successful and responsible.

Moving ahead

The journey toward AI-first is complex and there are many processes and experiences that will be transformed. Going AI-first means a transformation not just in technology but also in ethics, organizational culture, and human resource strategies. This then sets the stage for a more inclusive, adaptive, and forward-looking banking future — a live AI-first enterprise that continuously evolves and learns as new knowledge comes to light.